基于图像相邻像素相关性的 LSB 匹配隐写分析

摘要。:LSB匹配隐写是图像隐写分析中的重点研究问题。根据图像相邻像素的相关性,提出了一种新的隐写分析算法。通过图像复原算法计算出复原图像,利用高阶Markov链模型分别对待检测图像和复原图像建模,根据LSB匹配隐写对高阶Markov链模型经验矩阵的影响,提取复原图像和待检测图像的统计特征组合成新的27维特征向量对支持向量机进行训练。实验表明提出的算法对LSB匹配隐写有较好的分析效果,特别在嵌入率低的情况下,算法具有较好的分析能力。...

Full description

Saved in:
Bibliographic Details
Published in计算机应用研究 Vol. 31; no. 3; pp. 846 - 849
Main Author 吴松 张敏情 雷雨
Format Journal Article
LanguageChinese
Published Key Laboratory of Network & Information Security under the Chinese Armed Police Force of Dept. of Electronic,Engineering College of Chinese Armied Police Force,Xi'an 710086,China%Key Laboratory of Network & Information Security under the Chinese Armed Police Force of Dept. of Electronic,Engineering College of Chinese Armied Police Force,Xi'an 710086,China 2014
Institute of Information Security,Engineering College of Chinese Armied Poli
Subjects
Online AccessGet full text
ISSN1001-3695
DOI10.3969/j.issn.1001-3695.2014.03.050

Cover

More Information
Summary:摘要。:LSB匹配隐写是图像隐写分析中的重点研究问题。根据图像相邻像素的相关性,提出了一种新的隐写分析算法。通过图像复原算法计算出复原图像,利用高阶Markov链模型分别对待检测图像和复原图像建模,根据LSB匹配隐写对高阶Markov链模型经验矩阵的影响,提取复原图像和待检测图像的统计特征组合成新的27维特征向量对支持向量机进行训练。实验表明提出的算法对LSB匹配隐写有较好的分析效果,特别在嵌入率低的情况下,算法具有较好的分析能力。
Bibliography:51-1196/TP
WU Songa, ZHANG Min-qing, LEI Yua ( a. Key Laboratory of Network & Information Security under the Chinese Armed Police Force, Dept. of Electronic. b. Institute of Information Se- curity, Engineering College of Chinese Armied Police Force, Xi ' an 710086, China)
LSB matching; image restoration; high-order Markov chain; combination feature; low embedding rate
Detection of LSB matching steganography is an important research subject in image steganalysis. This paper pro- posed a new steganalysis algorithm based on the neighborhood correlation of pixels. It calculated restoration image through im- age restoration algorithms, and then modeled the restoration image and detection image with high order Markov chain modeled. It extracted statistical features of the restoration image and the detection image according to the effect of LSB matching on em- pirical matrix, then trained the vector support machine with the new combined 27 dimension features. Experiment results show the Proposed algorithm has a great e
ISSN:1001-3695
DOI:10.3969/j.issn.1001-3695.2014.03.050